XiaomiMiMo-V2.6-Pro
MiMo-V2.6-Pro is a Xiaomi open-weights reasoning model, released 21 Sept 2026. Its best configuration ranks #47 of 756 on the BenchLeader Index at 64.8 ±5.3, in the upper half. The ± is the point: 134 other configurations score within that range, so they and this one cannot be told apart on quality alone — price and speed are what separate them. It scores highest in composite (83) and lowest in agents & tools (57). At $0.544 per million tokens blended it is mid-priced. Output speed of 46 tokens per second puts it in the slowest quarter, with a first answer in 48.0 s. Last measured 2 Oct 2026.
- Blended price
- $0.544/M
- $0.435 in · $0.870 out
- Output speed
- 46 tok/s
- measured by Artificial Analysis
- First answer
- 48 s
- first token 3.45 s
- Context
- 1.0M
- Full answer
- 59 s
- median, reasoning included
- Cost per run
- $0.133
- one full Intelligence Index run
- Released
- 21 Sept 2026
- Overall index65
- Reasoning80
- Coding62
- Agents & tools57
- Maths67
- Knowledge66
- Human preference68
- Multimodal60
- Long context69
- Composite83
Versions
Xiaomi has shipped 2 models under this name. Each is ranked on its own results; a newer version often has fewer results so far, which holds its index nearer the average until more arrive.
| Model | Released | Index | Rank |
|---|---|---|---|
| MiMo-V2.6-Prothis page | 21 Sept 2026 | 64.8 | #47 |
| MiMo-V2.5-Prothinking | 23 Apr 2026 | 57.2 | #175 |
How its index has moved
23 Sept 2026 to 5 Oct 2026The index is recomputed from scratch every day, so a line moves when a new benchmark result lands, when a publisher revises a score, or when the models it is normalised against change. Early movement usually means the score is still settling.
Benchmark results
One column per reasoning effort. Rank is among every configuration of every model on that benchmark. Hover a score for the run it came from.
Reasoning
| Benchmark | not stated | Source | Trend |
|---|---|---|---|
| LMArena Hard Prompts | 1506#22 | LMArena | |
| Humanity's Last Exam (AA)not in index | 49.4%#26 | Artificial Analysis | |
| CritPt | 26.6%#35 | Artificial Analysis | |
| MysteryMechanismnot in index | 15.3%#20 | Vals AI |
Coding
| Benchmark | not stated | Source | Trend |
|---|---|---|---|
| SciCode | 60.9%#7 | SciCode | |
| LMArena Coding | 1540#10 | LMArena | |
| LMArena WebDev | 1618#23 | LMArena | |
| SciCode (AA)not in index | 60.9%#7 | Artificial Analysis | |
| IOI | 39.3%#30 | Vals AI | |
| Code Migrationnot in index | 43.0%#22 | Vals AI | |
| ProgramBenchnot in index | 0.5%#13 | Vals AI | |
| Vibe Code Bench v1.1not in index | 85.2%#12 | Vals AI | |
| ALE-Benchnot in index | 1158.1#41 | Epoch AI Benchmarking Hub | |
| Terminal-Bench 4.0 (AA)not in index | 34.9%#32 | Artificial Analysis |
Agents & tools
| Benchmark | not stated | Source | Trend |
|---|---|---|---|
| APEX-Agents | 59.5%#14 | Mercor | |
| Terminal-Bench 2.1 (Vals) | 67.8%#32 | Vals AI | |
| GDPval (AA)not in index | 59.3%#11 | Artificial Analysis | |
| CyberBenchnot in index | 72.9%#11 | Vals AI | |
| Finance Agent v2not in index | 57.3%#13 | Vals AI | |
| Harvey's Legal Agent Benchmarknot in index | 10.8%#12 | Vals AI | |
| Legal Research Benchnot in index | 47.1%#12 | Vals AI | |
| SREBenchnot in index | 3.0%#12 | Vals AI | |
| Tax Agent Benchnot in index | 64.9%#18 | Vals AI | |
| Terminal-Bench 4.0 (Vals)not in index | 31.3%#13 | Vals AI | |
| Terminal-Bench Sciencenot in index | 2.9%#23 | Vals AI | |
| AutomationBenchnot in index | 58.6%#23 | Artificial Analysis | |
| GDP.pdfnot in index | 19.2%#31 | Artificial Analysis | |
| AA-Briefcasenot in index | 1516#21 | Artificial Analysis |
Maths
| Benchmark | not stated | Source | Trend |
|---|---|---|---|
| ProofBench | 70.0%#14 | Vals AI | |
| LMArena Maths | 1487#30 | LMArena |
Knowledge
| Benchmark | not stated | Source | Trend |
|---|---|---|---|
| AA-Omniscience | 8.4#92 | Artificial Analysis | |
| Excel Modeling Benchmarknot in index | 62.9%#27 | Vals AI | |
| MedCodenot in index | 45.0%#37 | Vals AI | |
| MedScribenot in index | 88.3%#11 | Vals AI | |
| Public Benefits Benchnot in index | 68.9%#6 | Vals AI | |
| AA-Omniscience: accuracynot in index | 34.9%#145 | Artificial Analysis | |
| AA-Omniscience: non-hallucinationnot in index | 59.4%#74 | Artificial Analysis |
Instruction following
| Benchmark | not stated | Source | Trend |
|---|---|---|---|
| LMArena Instruction Followingnot in index | 1474#29 | LMArena |
Human preference
| Benchmark | not stated | Source | Trend |
|---|---|---|---|
| LMArena Text | 1480#26 | LMArena | |
| LMArena Creative Writingnot in index | 1446#47 | LMArena | |
| LMArena Multi-turnnot in index | 1456#81 | LMArena | |
| LMArena Longer Queriesnot in index | 1495#21 | LMArena |
Multimodal
| Benchmark | not stated | Source | Trend |
|---|---|---|---|
| LMArena Vision | 1264#49 | LMArena | |
| SAGEnot in index | 45.0%#40 | Vals AI |
Long context
| Benchmark | not stated | Source | Trend |
|---|---|---|---|
| AA-LCR | 86.3%#3 | Artificial Analysis | |
| MLCRnot in index | 18.3%#20 | Artificial Analysis |
Composite
| Benchmark | not stated | Source | Trend |
|---|---|---|---|
| AA Intelligence Index | 46.3#31 | Artificial Analysis | |
| Vals Indexnot in index | 55.2#12 | Vals AI |
Where it wins
Benchmarks where this configuration ranks in the top five of every configuration measured.
Where to run it
Every provider serving this model through OpenRouter, with throughput and first-token latency measured on live traffic over the last 30 minutes and each provider’s own price. Purple marks the best in each column.
| Provider | Speed | First token | Input $/M | Output $/M | Context | Quantisation |
|---|---|---|---|---|---|---|
| NovitaAI | 30 tok/s | 4.13 s | $0.435 | $0.870 | 1.0M | fp8 |
| Xiaomi | 30 tok/s | 3.45 s | $0.435 | $0.870 | 1.0M | fp8 |
| GMICloud | 27 tok/s | 6.48 s | $0.435 | $0.870 | 1.1M | bf16 |
| DeepInfra | 26 tok/s | 2.07 s | $0.430 | $0.870 | 1.0M | fp8 |
What a task costs
Estimates from list price, output speed and time to first answer for the best configuration. “With caching” assumes three-quarters of the input is served from the prompt cache at $0.004 per 1M. Reasoning tokens are not modelled.
| Workload | Tokens in / out | Cost | With caching | Time |
|---|---|---|---|---|
| Chat reply | 400 / 300 | $0.0004 | $0.0003 | 54.6 s |
| Summarise a 30-page report | 12,000 / 600 | $0.0057 | $0.0019 | 1.0 min |
| Code edit | 6,000 / 1,500 | $0.0039 | $0.0020 | 1.3 min |
| Agentic coding session | 60,000 / 4,000 | $0.030 | $0.010 | 2.3 min |
| Structured extraction | 2,000 / 200 | $0.0010 | $0.0004 | 52.4 s |
See also
Data as of 5 Oct 2026. Compare with another model.
Cite as: BenchLeader, “MiMo-V2.6-Pro: benchmarks, pricing, speed and rank”, https://www.benchleader.com/models/mimo-v2-6-pro, data as of 5 Oct 2026.